Vol. 19: 297–305, 2013 AQUATIC BIOLOGY Published online November 11 doi: 10.3354/ab00539 Aquat Biol

Shell morphology and genetic characteristics of () spp. (: ) from the waters adjacent to Japan, Korea and San Francisco Bay

Masato Owada1,*, Yoshitaro Kasai2, Shin’ichi Sato3, Jae-Sang Hong4

1Department of Biological Sciences, Kanagawa University, 2946 Tsuchiya, Hiratsuka 259-1293, Japan 2Department of Earth Sciences, Graduate School of Science, Tohoku University, 6-3 Aoba, Aramaki, Aoba-ku, Sendai 980-8578, Japan 3The Tohoku University Museum, 6-3 Aoba, Aramaki, Aoba-ku, Sendai 980-8578, Japan 4Department of Oceanography, Inha University, Incheon 402-751, South Korea

ABSTRACT: Corbula (Potamocorbula) spp. are natively distributed in the waters adjacent to Japan and Korea. These populations are distinguished by 6 morphological characteristics. How- ever, they have high geographic variation, and it is quite difficult to distinguish among them. In the present study, we investigated the geographic distribution of the morphological and genetic characteristics of Corbula (Potamocorbula) spp. collected from the waters adjacent to Japan, Korea and San Francisco Bay, USA. Shell morphology was inferred through canonical discrimi- nant analysis (CDA) of 8 shell characteristics, and genetic characteristics were analysed using the DNA sequences of mitochondrial cytochrome c oxidase subunit 1 (CO1) and nuclear internal tran- scribed spacer 2 (ITS2) regions. The results of CDA showed that the samples from Ganghwa-do in western Korea could be clearly separated from those of all the other localities. The molecular phy- logenetic tree and haplotype network that were constructed from the CO1 and ITS2 regions, respectively, did not show clear groupings among the collected samples, except for 1 group com- posed of samples from Ganghwa-do. In addition, no differences in CO1 amino acid sequences could be found among any of the samples examined, including the Ganghwa-do samples. These facts suggest that Corbula (Potamocorbula) spp. distributed in the waters adjacent to Japan and Korea were separated into 2 groups. However, these 2 groups were genetically closely related, and it seems appropriate to treat them as 2 distinct . The classification of the subgenus Potamocorbula should be re-examined at the species level.

KEY WORDS: · Bivalvia · Corbulidae · Potamocorbula · Phylogeography · Classification · Mitochondrial cytochrome c oxidase subunit 1 · Nuclear internal transcribed spacer 2

Resale or republication not permitted without written consent of the publisher

INTRODUCTION East Asia (Carlton et al. 1990, Nichols et al. 1990). It is known that the waters adjacent to Japan and Populations of the Asian (overbite) clam, Corbula Korea in the Northwest Pacific are inhabited by at (Potamocorbula) amurensis Schrenck, 1861, in San least 4 species of the subgenus Potamocorbula Habe, Francisco Bay, USA, have markedly increased since 1955. These species are distinguishable from each 1986. They were introduced to the region through other by 6 morphological characteristics: shell size, ballast waters from a cargo ship that had stopped in shell outline, shell inflation, shell thickness, posterior

*Email: [email protected] © Inter-Research 2013 · www.int-res.com 298 Aquat Biol 19: 297–305, 2013

margin and position of the (Zhuang & Cai 1983, DNA. However, they did not simultaneously discuss Habe 1994, Horikoshi & Okamoto 1994, Tsuchida & the relationships between the morphological and Okamura 1997). However, Sato (2002, 2005) and Sato genetic characteristics. In the present study, we & Kanazawa (2004) reported that there were high investigated the geographic distribution of the mor- geographic variations in these morphological charac- phological and genetic characteristics of the popula- teristics, making it challenging to distinguish among tions from the waters adjacent to Japan, Korea and these species. San Francisco Bay through canonical discriminant Corbula (Potamocorbula) amurensis is regarded as analysis (CDA) of 8 shell characteristics and through an in San Francisco Bay and may the DNA sequences of mitochondrial cytochrome c have a significant impact on the estuarine ecosystem oxidase subunit 1 (CO1) and nuclear internal tran- of the bay (Carlton 1999, Coan 2002, Kimmerer 2002, scribed spacer 2 (ITS2) regions. Greene et al. 2011). In addition, fossil specimens of this species have been used as indicators of a cool cli- mate or brackish water during reconstructions of the MATERIALS AND METHODS paleoenvironments of the Late Pleistocene and Holocene in Japan (Matsushima 1984, Habe 1994). Sampling However, these may not be accurate until the rela- tionships between the geographic distribution and Corbula (Potamocorbula) spp. were collected from morphological characteristics of Corbula (Potamocor- 8 localities shown in Fig. 1: Rokkaku River Estuary, bula) spp. are fully understood at both population Saga Prefecture, Ariake Sea, western Japan (N = 25 and species levels. To clarify the relationships among individuals collected on 12 July 1998 and 23 Novem- members of Corbula (Potamocorbula) spp., it will be ber 2008); Dong River Estuary, Suncheon-si, south- necessary to investigate the morphological and ern Korea (N = 30, 20 August 2011); Mangyeong genetic characteristics of populations from various River Estuary, Gunsan-si, western Korea (N = 30, 3 localities. There have been several discussions of the May 2000 and 11 July 2008); Ganghwa-do, intertidal genetic characteristics of Corbula (Potamocorbula) muddy sand flat, western Korea (N = 30, 8 May 1998 spp. For example, Duda (1994) reported that C. (P.) and 18 August 2011); Maeho Lake, Yangyang-gun, amurensis from San Francisco Bay, California, shows eastern Korea (N = 30, 16 July 2008); Hinuma Lake, high genetic variation. Further, Smith et al. (2012) Ibaraki Prefecture, central Japan (N = 30, 31 May indicated that C. (P.) amurensis from San Francisco 1998 and 30 July 2011); Obuchi Swamp, Aomori Pre- Bay and the Ariake Sea, western Japan, can be iden- fecture, northern Japan (N = 46, 3 September 1998 tified using the same genetic marker, 18S ribosomal and 27 July 2009); and San Francisco Bay, California,

45° 45° N N

USA Os Pacific Japan / East Sea Sf Seoul Ml Gd San Francisco South Japan Yellow Mr Hl Sea Korea 35° Tokyo 35° Dr Pacific Ocean Rr 200 km 200 km

130°E 140° 130°W 120° Fig. 1. Localities where the samples were collected. Rr: Rokkaku River Estuary (33°11’ N, 130°12’ E); Dr: Dong River Estuary (34° 51’ N, 127° 30’ E); Mr: Mangyeong River Estuary (35° 51’ N, 126° 40’ E); Gd: Ganghwa-do (37° 36’ N, 126° 22’ E); Ml: Maeho Lake (37° 56’ N, 128° 46’ E); Hl: Hinuma Lake (36° 16’ N, 140° 30’ E); Os: Obuchi Swamp (40° 57’ N, 141° 22’ E); Sf: San Francisco Bay (37° 27’ N, 122° 1’ W) Owada et al.: Morphological and genetic characteristics of Corbula (Potamocorbula) spp. 299

USA (N =30, 9 February 2000 and 9 November 2010). Polymerase chain reaction (PCR) In addition, C. (P.) fasciata and Solidicorbula ery- throdon, which were used as outgroups in the ana- CO1 and ITS2 regions were amplified using Pre- lysis of genetic variation, were collected from Bohol mix Taq (Takara) and a PCR Thermal Cycler PER- Island in the Philippines (9°44’ N, 123°53’ E) on 8 SONAL (Takara). The following primers were newly June 2007 and from Sagami Bay in Japan (35°17’ N, designed for the present study: 5′-TCW ACA AAT 139°20’ E) on 7 September 2011, respectively. The CAY AAA GAT ATT GG-3′ (forward) and 5′-ACY soft tissue of living was preserved in 100% TCM GGR TGM CCA AAA AAT CA-3′ (reverse) for ethanol, and DNA was extracted from the foot or the CO1 region, and 5′-ATT GCA GGA CAC ATT adductor muscle using the DNeasy Blood & Tissue GAA CAT CG-3′ (forward) and 5′-GTT ACT GAG Kit (Qiagen). The shell and soft tissue were num- GGA ATC CTT GTT A-3′ (reverse) for the ITS2 bered in each individual. All voucher samples were region. The conditions for PCR amplification were as deposited in the Tohoku University Museum (TUMC follows: denaturation at 94°C for 30 s, annealing at Nos. 111200 to 111207). 54°C for 30 s (CO1) and 55°C for 30 s (ITS2) and extension at 72°C for 60 s. These steps were repeated 30 times. The PCR products were purified using Morphological analysis ExoSAP-IT (USB), and cycle-sequencing reactions were performed using BigDye Terminator Version Eight shell characteristics (in mm): shell length 3.1 Cycle Sequencing Kit (ABI PRISM). The se - (L), shell breadth (B), shell height of left valve (HL), quences were compiled using a 3130 Genetic Ana- shell height of right valve (HR), thickness of shell lyzer (ABI PRISM) from both the 5′ and 3′ directions. margin (T), height of umbo (U), posterior shell These sequences were registered in the DNA Data length (PL) and upper shell height (UH) were meas- Bank of Japan (CO1: AB740093 to AB740137 and ured on each individual (Fig. 2). For measurement AB827932 to AB827935, ITS2: AB740177 to AB740218 of L, HR, U, PL and UH, the outside of the right shell and AB827936 to AB827938). valve was photographed with a digital camera, and then each characteristic was measured using image analysis software, ImageJ ver. 1.45s. Further, L, B, Genetic analysis HL, HR and T were measured using a digital slide caliper (accuracy ± 0.01 mm). L and HR were meas- In the analysis of genetic variation in the CO1 re- ured using both methods, but there were no signifi- gion, the sequences of Varicorbula dissimilis (AF cant differences (p > 0.05) between them. Then, to 120669), Solidicorbula erythrodon (AB740137), and standardize the variability for size, all characteristics Corbula (Potamocorbula) fasciata (AB740136) were were log-transformed (using base 10 logs), and the included as outgroups. After alignment using Clustal 7 characteristics excluding L were divided by the W Version 1.8 (Thompson et al. 1997), sites that log L. CDA was tested with the standardized 7 char- contained gaps or those that could not be confirmed acteristics using SPSS (Ver. 16.0). by homology were removed from the analysis. A mo- lecular phylogenetic tree was reconstructed using maximum-likelihood (ML) and Bayesian methods, re- U spectively. The ML method was performed by PAUP B Version 4.0*b10 (Swofford 2002). The tree search UH HL algorithm was a likelihood ratchet algorithm (2000 PL or HR times) (Vos 2003), and the model for the ML method was selected using Modeltest Version 3.7 (Posada & Crandall 1998). A bootstrap test was conducted 1000 T times. The Bayesian method was performed by Mr- Bayes Version 3.1.2 (Ronquist & Huelsenbeck 2003). L This program was run for 5 000 000 gen erations, with Fig. 2. Corbula (Potamocorbula) spp. Eight characteristics of sampling every 1000th generation. The model for the left and right shell valves and shell breadths measured on Bayesian method was determined using Mrmodeltest shells. : shell length; : shell breadth; : shell height of L B HL Version 2.3 (Posada & Buckley 2004). left valve; HR: shell height of right valve; T: thickness of shell margin; U: height of umbo; PL: posterior shell length; In the analysis of genetic variation in the ITS2 UH: upper shell height region, the sequences were aligned using Clustal W 300 Aquat Biol 19: 297–305, 2013

Version 1.8 (Thompson et al. 1997). Sites that con- largely affect Canonical Variate 1 were shell height tained gaps and those that could not be confirmed by and height of umbo (Table 2). Namely, individuals homology were removed from the analysis. Further, from Ganghwa-do have smaller shell heights and a haplotype network was built using TCS Version umbos rather than those from the other 7 localities 1.21 (Clement et al. 2000). (Fig. 3). The other 7 samples beside Ganghwa-do cannot be clearly separated from each other (Fig. 4). However, RESULTS individuals from Maeho Lake and Hinuma Lake have similar shell morphology, and those from Rokkaku Canonical discriminant analysis River Estuary, Mangyeong River Estuary and San Francisco Bay were also of similar shell shape Shells of representative individuals collected from (Fig. 4). The samples from Dong River Estuary and each locality are shown in Fig. 3. According to the Obuchi Swamp were slightly different from the other results of CDA, individuals from Ganghwa-do can be 5 samples, but some individuals overlapped morpho- clearly separated from those of all the other localities logically (Fig. 4). In conclusion, though these 7 sam- by the 7 standard morphological characteristics (Fig. 4). ples have a characteristic shell morphology, they The proportions of Canonical Variates 1 and 2 were cannot be distinguished because of the high morpho- 84.6 and 8.3%, respectively (Table 1). The centroid of logical variability. Canonical Variate 1 in individuals from Ganghwa-do Among the samples from each locality, the percent was less than −6, but those from the other 7 localities of correct cases in multivariate discriminant analysis were all more than −2. The shell characteristics that ranged between 100 and 60.0% (Table 3). All indi-

Fig. 3. Corbula (Potamocorbula) spp. Representative individuals collected from each locality. For each locality the upper 2 rows display the right valve and the lower 2 rows display the left valve. Scale bars: 5 mm. Rr: Rokkaku River Estuary; Dr: Dong River Estuary; Mr: Mangyeong River Estuary; Gd: Ganghwa-do; Ml: Maeho Lake; Hl: Hinuma Lake; Os: Obuchi Swamp; Sf: San Francisco Bay Owada et al.: Morphological and genetic characteristics of Corbula (Potamocorbula) spp. 301

10 Rr Dr Mr Gd Ml Hl Os Sf 5 Group centroid

0 Canonical variate 2 Canonical variate

-5

-10 -10 -5 0 5 10 Canonical variate 1

Fig. 4. Corbula (Potamocorbula) spp. Two-dimensional scattergrams of Corbula (Potamocorbula) spp. from 8 localities ob- tained with canonical discriminant analysis. The values of Canonical Variates 1 and 2 for each individual and the centroids for individuals from each locality are shown. Rr: Rokkaku River Estuary; Dr: Dong River Estuary; Mr: Mangyeong River Estuary; Gd: Ganghwa-do; Ml: Maeho Lake; Hl: Hinuma Lake; Os: Obuchi Swamp; Sf: San Francisco Bay

Table 1. Corbula (Potamocorbula) spp. Standardized coeffi- cients of 7 characteristics, eigenvalue and proportion of Canonical Variates 1 and 2 for 8 groups. L: shell length; B: Table 2. Corbula (Potamocorbula) spp. Centroids of Canoni- shell breadth; HL: shell height of left valve; HR: shell height cal Variates 1 and 2 for individuals from 8 localities. Rr: of right valve; T: thickness of shell margin; U: height of Rokkaku River Estuary; Dr: Dong River Estuary; Mr: umbo; PL: posterior shell length; UH: upper shell height Mangyeong River Estuary; Gd: Ganghwa-do; Ml: Maeho Lake; Hl: Hinuma Lake; Os: Obuchi Swamp; Sf: San Francisco Bay Characteristics Standardized coefficients of Canonical Variates: 12Locality Centroid of Canonical Variate: 12 log B/log L −0.248 1.115 log HL/log L 0.385 0.255 Rr 1.589 −0.440 log HR/log L 0.361 −0.475 Dr −1.377 0.702 log T/log L 0.357 −0.199 Mr 1.496 −0.809 log U/log L 0.451 0.226 Gd −6.914 0.651 log PL/log L −0.200 0.227 Ml 3.127 1.317 log UH/log L 0.095 −0.853 Hl 4.008 0.752 Eigenvalue 11.314 1.108 Os −1.822 −0.258 Proportion 84.6 8.3 Sf −0.077 −1.980 302 Aquat Biol 19: 297–305, 2013

Table 3. Corbula (Potamocorbula) spp. Discrimination scores using multivariate discriminant analysis among individuals from 8 localities. Rr: Rokkaku River Estuary; Dr: Dong River Estuary; Mr: Mangyeong River Estuary; Gd: Ganghwa-do; Ml: Maeho Lake; Hl: Hinuma Lake; Os: Obuchi Swamp; Sf: San Francisco Bay

No. of No. of Percent of Number of classified cases total cases correct cases correct cases Rr Dr Mr Gd Ml Hl Os Sf

Rr 25 19 76.0 19 1500000 Dr 30 27 90.0 0 27 210000 Mr 30 18 60.0 6 1 18 01103 Gd 30 30 100.0 0 0 0 30 0000 Ml 29 22 75.9 111022400 Hl 28 24 85.7 000042400 Os 20 19 95.0 010000190 Sf 30 20 66.7 108000120 Total 222 179 80.6 27 31 34 31 27 29 20 23

viduals from Ganghwa-do were classified Solidicorbula correctly, but 25.0% of the individuals from erythrodon the other 7 localities were classified incor- ● Varicorbula rectly (43 out of 172 ind.) (Table 3). Corbula (Potamocorbula) fasciata Gd01 dissimilis Gd08 Gd02 ● Gd04 Genetic variation of the CO1 region Gd05 Gd06 DNA sequences of 606 bp in length were Gd03 Gd09 compiled for 47 individuals. They were 5 ● Mr06 sequences from Rokkaku River Estuary, 4 Rr07 Dr02 from Dong River Estuary, 8 from Mangyeong Mr04

River Estuary, 8 from Ganghwa-do, 3 from ● Mr09 Maeho Lake, 4 from Hinuma Lake, 8 from Sf03 Sf05 Obuchi Swamp and 7 from San Francisco Hl02 Bay. After the alignment, a 594 bp sequence Hl03 0.1 substitutions/site Ml02 was used for the genetic analysis. Genetic Ml03 variation among the sequences was found Ml05 at 64 nucleotide sites (11%); however, the Rr06 Os01 sequences of some individuals were identi- Os02 cal. The sequences that were translated to Os03 Fig. 5. Corbula (Potamocorbula) Os04 spp. Molecular phylogenetic amino acids were the same in all the indi- Os05 tree on the basis of the DNA se- viduals from the 8 localities. There were Os06 quences of the mitochondrial 25 amino acid substitutions between the Os07 cytochrome c oxidase subunit 1 Hl01 region. A filled circle on a node ingroup and the outgroup Corbula (Pota- Hl04 indicates a bootstrap value mocorbula) fasciata. Os10 Dr01 >50% and a Bayesian posterior In the ML method, the K81uf+G model was Dr03 probability of >0.90 for the re- selected by Modeltest Version 3.7 and the Dr04 spective monophyletic clade. likelihood index (−ln L) was 2593.56183. In Mr02 Branch length was calculated Mr03 the Bayesian method, the GTR+G model was using the maximum-likelihood Mr05 method. Rr01 to 08: individuals selected by Mrmodeltest. A strict consen- Mr08 Mr10 from Rokkaku River Estuary; sus tree was computed from these 2 trees Sf02 Dr01 to 04: Dong River Estuary; (Fig. 5). In the strict consensus tree, all indi- Sf04 Mr02 to 10: Mangyeong River viduals from the 8 localities formed a mono- Sf12 Estuary; Gd01 to 09: Ganghwa- Sf13 do; Ml02 to 05: Maeho Lake; phyletic group and Corbula (Potamocorbula) Sf14 Hl01 to 04: Hinuma Lake; Os01 fasciata was identified as a sister group of the Rr01 to 10: Obuchi Swamp; Sf02 to Rr03 monophyletic group. The individuals from Rr08 13: San Francisco Bay Owada et al.: Morphological and genetic characteristics of Corbula (Potamocorbula) spp. 303

Ganghwa-do formed a monophyletic group, and it individuals. In the Ganghwa-do group, 3 haplotypes was a sister group of the other 7 localities. Some indi- were identified, and the highest divergence was viduals from Rokkaku River Estuary, Mangyeong 3 bp. In the group containing all the other individuals, River Estuary and San Francisco Bay formed mono- 7 haplotypes were identified, and the highest diver- phyletic groups (Fig. 5). gence was 4 bp. One haplotype was found in the samples from 6 investigated localities, except for those from Dong River Estuary and Ganghwa-do (Fig. 6). Genetic variation of the ITS2 region

Sequences of 384 to 393 bp in length were com- DISCUSSION piled for 45 individuals. They were 5 sequences from Rokkaku River Estuary, 3 from Dong River Estuary, 8 According to the criteria of Zhuang & Cai (1983), from Mangyeong River Estuary, 8 from Ganghwa-do, the samples from Rokkaku River Estuary, Mang - 4 from Maeho Lake, 4 from Hinuma Lake, 6 from yeong River Estuary and San Francisco Bay were Obuchi Swamp and 7 from San Francisco Bay. After identified as Corbula (Potamocorbula) amurensis, the alignment, a 377 bp sequence was used for ge- those from Ganghwa-do as a morphological type of netic analysis. Genetic variation among the sequences C. (P.) laevis Hinds, 1843, those from Maeho Lake was found at 16 sites (4%). and Hinuma Lake as C. (P.) ustulata (Reeve 1844) The haplotype network is shown in Fig. 6, and it and those from Dong River Estuary and Obuchi revealed 10 unique haplotypes forming 2 groups. Swamp as C. (P.) rubromuscula (Zhuang & Cai 1983). The highest divergence among them was 11 bp. One However, Horikoshi & Okamoto (1994) identified the group included only individuals from Ganghwa-do, samples from Rokkaku River Estuary as C. (P.) cf. and the other group was composed of the remaining laevis. Tsuchida & Okamura (1997) identified the

1 Mr Os 3 Dr

Dr Sf Sf 1 Gd 7 Rr Os 6 Os Mr Hl Ml

Sf Rr

Os 1818 6 Mr Gd Hl Ml

1 Os 1 Gd 1 Sf

Fig. 6. Corbula (Potamocorbula) spp. Haplotype network on the basis of the DNA sequences of the nuclear internal transcribed spacer 2 region. Numerals indicate the number of individuals in the samples. Rr: Rokkaku River Estuary; Dr: Dong River Estuary; Mr: Mangyeong River Estuary; Gd: Ganghwa-do; Ml: Maeho Lake; Hl: Hinuma Lake; Os: Obuchi Swamp; Sf: San Francisco Bay 304 Aquat Biol 19: 297–305, 2013

samples from Hinuma Lake as C. (P.) ustulata, in viduals from Mangyeong River Estuary and Rokkaku agreement with Zhuang & Cai (1983), though Huber River Estuary formed a clade with those from San (2010) defined this species as a junior synonym of Francisco Bay in the phylogenetic tree, and the sam- C.(P.) nimbosa Hanley, 1843. ples from all the investigated localities, except for On the basis of their classification, Corbula (Pota- those from Dong River Estuary and Ganghwa-do, mocorbula) spp. distributed in the waters adjacent to contained the same haplotype in the haplotype net- Japan and Korea are divided into at least 4 species. If work. These facts suggest that the artificially intro- the classification is accurate, then these groups could duced populations also belonged to the same his- be indicated through the analysis of morphological torically large population. Therefore, it may be that and molecular characteristics. However, the in vesti - the historical population was naturally or artificially gated samples did not form any species groups, separated and that the populations of the 7 locali- except for the group composed of samples from ties were recently formed. Further, the population Ganghwa-do. In the CDA results, samples from of Ganghwa-do may have separated from the his - Ganghwa-do were clearly separated from those of torical population earlier than those of the other the other 7 localities. Among the CO1 and ITS2 localities. sequences, which were used to analyze intra-species Nichols et al. (1990) and Nicolini & Penry (2000) relationships, the greatest difference was found reported that Corbula (Potamocorbula) amurensis is between the groups composed of samples from euryhaline and inhabits seawater with salinity rang- Ganghwa-do and those composed of the remaining ing from 5 to 25‰. In addition, the populations of samples. This difference was larger than the largest Rokkaku River Estuary, Mangyeong River Estuary difference among the samples from the other 7 local- and San Francisco Bay have been commonly shown ities. Therefore, the samples from Ganghwa-do can to increase quickly after sudden environmental be distinguished from those of the other localities changes such as reclamation (Takayasu et al. 2000, on the basis of both morphological and genetic char- Sato et al. 2001, Sato 2002, 2005, Sato & Azuma 2002, acteristics. Furthermore, the samples from Ganghwa- Sato & Kanazawa 2004). These facts could cause high do inhabited the intertidal zone of a muddy sand variations in morphological characteristics. Although flat at an extremely high population density (maxi- Corbula (Potamocorbula) spp. have been treated as mum density >150 000 ind. m−2). This habitat is quite indicators of a cool climate or brackish water in the different from those of the other samples, which past, the results of the present study do not support mainly inhabited the subtidal zone of mud flats or this. The smaller shells of the Ganghwa-do popula- sand coasts. Therefore, reproductive isolation may tion might indicate a high population density, but it is have occurred between the Ganghwa-do population not clear whether genetic or environmental factors and the populations from the other localities. On the caused the shell reduction. Therefore, Corbula (Pota- other hand, no differences in CO1 amino acid mocorbula) spp. may not be used as a paleoenviron- sequences, which has been frequently applied as a mental indicator. criterion for differences between species, were iden- The present study showed that Corbula (Pota- tified among any of the investigated samples. This mocorbula) spp. distributed in the waters adjacent to suggests that the reproductive isolation occurred rela- Japan and Korea are divided into 2 groups on the tively recently. basis of morphological and genetic characteristics. Although the populations of Dong River Estuary, The 2 groups are closely related genetically; how- Mangyeong River Estuary, Maeho Lake and Obuchi ever, reproductive isolation between them is pos - Swamp are native to these places, we found no sible. Hence, currently, it seems appropriate to population-specific substitutions in CO1 and ITS2 treat them as at least 2 distinct species. The classifi- regions. Further, the geographical distance between cation of the subgenus Potamocorbula should be re- these localities would make it difficult for the popula- examined at the species level. tions inhabiting them to exchange genes. Thus, it Hallan & Willan (2010) described 2 species of may be concluded that these populations originated genus Lentidium from northern Australia. Lentidium from a common, historically large population. In con- conspicuously resembles the subgenus Potamo - trast, it is known that the populations of Rokkaku corbula in shell morphology. Thus, Lentidium and River Estuary, Hinuma Lake and San Francisco Bay Potamocorbula could be closely related, but further were artificially introduced from other localities information on both morphological and genetic char- (Carlton et al. 1990, Nichols et al. 1990, Horikoshi & acteristics is needed to investigate the relationship Okamoto 1994, Tsutsumi et al. 1997). Some indi - between the 2 genera (cf. Hallan et al. 2013). Owada et al.: Morphological and genetic characteristics of Corbula (Potamocorbula) spp. 305

Acknowledgements. We deeply thank Emr. Prof. I. Hayami, of a former community. Mar Ecol Prog Ser 66:95−101 Prof. K. Kanazawa and Prof. A. Hino (Kanagawa University) Nicolini MH, Penry DL (2000) Spawning, fertilization, and for their support of this study. We are indebted to Prof. T. larval development of (Mol- Yamaguchi (Kanagawa University) and Prof. J. T. Carlton lusca: Bivalvia) from San Francisco Bay, California. Pac (Williams College) for valuable comments, to Dr. R. Naka - Sci 54:377−388 shima (AIST) for fruitful discussions and to Drs T. Haga Posada D, Buckley TR (2004) Model selection and model (JAMSTEC) and J. K. Thompson (U.S. Geological Survey) averaging in phylogenetics: advantages of the AIC and for donating the samples and providing valuable comments. Bayesian approaches over likelihood ratio tests. Syst Biol We are also grateful to 2 anonymous reviewers for useful 53: 793−808 comments. We acknowledge the Japan/Korea Tidal-Flats Posada D, Crandall KA (1998) Modeltest: testing the model Joint Survey Group for their help in collecting the samples. of DNA substitution. Bioinformatics 14: 817−818 This work was supported by a Grant-in-Aids for Scientific Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian Research from the Japan Society for the Promotion of phylogenetic inference under mixed models. Bioinfor- Science (19650219 and 21500861 to S.S.) and a grant from matics 19: 1572−1574 the National Institute of Biological Resources, funded by Sato S (2002) Faunal response of bivalves and gastropods the Ministry of Environment of the Republic of Korea (NIBR of large environmental disturbances caused by the con- No. 2013-02-001). struction of dyke for reclamation. Jpn J Benthol 57: 106−118 (in Japanese with English abstract) Sato S (2005) Faunal response of bivalves to large envi- LITERATURE CITED ronmental disturbances caused by huge reclamation projects in Japan and Korea—a comparison with the Carlton JT (1999) Molluscan invasions in marine and estuar- fossil bivalve assemblages. Aquabiology 159: 356−361 (in ine communities. Malacologia 41: 439−454 Japanese) Carlton JT, Thompson JK, Schemel LE, Nichols FH (1990) Sato S, Azuma M (2002) Ecological and paleoecological Remarkable invasion of San Francisco Bay (California, implications of the rapid increase and decrease of an USA) by the Asian clam Potamocorbula amurensis. I. introduced bivalve Potamocorbula sp. after the con - Introduction and dispersal. Mar Ecol Prog Ser 66: 81−94 struction of a reclamation dike in Isahaya Bay, western Clement M, Posada D, Crandall K (2000) TCS: a computer Kyushu, Japan. Palaeogeogr Palaeoclimatol Palaeoecol program to estimate gene genealogies. Mol Ecol 9: 185: 369−378 1657−1659 Sato S, Kanazawa T (2004) Faunal change of bivalves in Ari- Coan EV (2002) The eastern Pacific recent species of the ake Sea after the construction of the dike for reclamation Corbulidae (Bivalvia). Malacologia 44: 47−105 in Isahaya Bay, western Kyushu, Japan. Kaseki 76: 90−99 Duda TF (1994) Genetic population structure of the recently (in Japanese with English abstract) introduced Asian clam, Potamocorbula amurensis, in San Sato S, Azuma M, Kondo H, Nishinokubi H (2001) Temporal Francisco Bay. Mar Biol 119:235−241 changes of bivalves and gastropods in Isahaya Bay shut Greene VE, Sullivan LJ, Thompson JK, Kimmerer WJ (2011) off from Ariake Sea, western Kyushu, Japan. Quat Res Grazing impact of the invasive clam Corbula amurensis (Tokyo) 40: 85−95 (in Japanese with English abstract) on the microplankton assemblage of the northern San Smith KF, Wood SA, Mountfort DO, Cary SC (2012) Devel- Francisco Estuary. Mar Ecol Prog Ser 431:183−193 opment of a real-time PCR assay for the detection of the Habe T (1994) Potamocorbula amurensis. In: Fisheries invasive clam, Corbula amurensis, in environmental Agency (ed) Basic data of rare wild aquatic organisms samples. J Exp Mar Biol Ecol 412: 52−57 in Japan. I. Japan Fisheries Resource Conservation Asso- Swofford DL (2002) PAUP*: phylogenetic analysis using ciation, Tokyo, p 79−80 (in Japanese) parsimony (and other methods), Version 4.0b10. Sinauer Hallan A, Willan R (2010) Two new species of Lentidium Associates, Sunderland, MA (: Corbulidae) from tropical northern Australia: Takayasu K, Tanaka H, Sato S (2000) Paleoenvironmental remarkable fresh/fluviatile to brackish-water bivalves. meanings of the brackish-water clam Potamocorbula Molluscan Res 31: 143−153 swarm beds in the sediment core SJ 96 from Lake Shinji, Hallan A, Colgan DJ, Anderson LC, García A, Chivas AR southwest Japan—in comparison with the occurrence (2013) A single origin for the limnetic-euryhaline taxa in in San Francisco Bay, California, USA. Geoscience Rept the Corbulidae (Bivalvia). Zool Scr 42:278–287 Shimane Univ 19: 37−45 (in Japanese with English Horikoshi M, Okamoto M (1994) An immigrant bivalve Pota- abstract) mocorbula cf. laevis (Hinds) recently propagated at the Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Hig- head of Ariake Bay, Kyushu, Japan. Chiribotan 24: 77−83 gins DG (1997) The Clustal X windows interface: flexible (in Japanese) strategies for multiple sequence alignment aided by Huber M (2010) Compendium of bivalves. ConchBooks, quality analysis tools. Nucleic Acids Res 25:4876−4882 Hackenheim Tsuchida E, Okamura S (1997) Some commercial mollusks Kimmerer WJ (2002) Effects of freshwater flow on abun- collected at Shanghai and Xiamen, China. Chiribotan 28: dance of estuarine organisms: physical effects or trophic 39−44 (in Japanese) linkages? Mar Ecol Prog Ser 243: 39−55 Tsutsumi T, Nishimura K, Okamoto M (1997) Potamocorbula Matsushima Y (1984) Shallow marine molluscan assem- amurensis occurred at Hi-numa Swamp in Ibaraki Pre- blages of postglacial period in the Japanese Islands—its fecture, eastern Japan. Chiribotan 28: 11−12 (in Japan- historical and geographical changes induced by the ese) environmental changes. Bull Kanagawa Prefect Mus 15: Vos RA (2003) Accelerated likelihood surface exploration: 37−109 (in Japanese with German and English abstract) the likelihood ratchet. Syst Biol 52:368−373 Nichols FH, Thompson JK, Schemel LE (1990) Remarkable Zhuang Q, Cai Y (1983) Studies on the Corbulidae (Bivalvia) invasion of San Francisco Bay (California, USA) by the off Chinese coasts. Trans Chinese Soc Malacol 1: 57−68 Asian clam Potamocorbula amurensis. II. Displacement (in Chinese with English abstract)

Editorial responsibility: Nikolaos Schizas, Submitted: January 2, 2013; Accepted: September 11, 2013 Mayagüez, Puerto Rico, USA Proofs received from author(s): November 1, 2013